---
title: "Ch06 brain in the vat content"
description: "Philosophers of mind use the brain in the vat to name a cognitive system modeled without a body and without a world: everything the system needs to think is…"
url: "https://wiki.cybernetic-futures.org/punch-and-judy/part-01-inside-the-model/ch06-brain-in-the-vat/ch06-brain-in-the-vat-content"
modified: "2026-11-12T15:26:25.887Z"
language: "en"
site: "Cybernetic Futures"
---

# The Brain in the Vat

Philosophers of mind use the brain in the vat to name a cognitive system modeled without a body and without a world: everything the system needs to think is inside it, and the environment enters only as signal.[^ch06-vat] The phrase describes, with uncomfortable accuracy, the shape of a diagram we draw several times a week.

## The diagram stops at the neck

We draw the model as a nervous system: channels, balances between them, a function that scans an environment, and an identity placed at the top of the page. The vocabulary is neurological and so is the picture, and both stop where a neck would be.[^ch06-teaching] Beer's model does have a body: the body of the organization, whose organs are the operating units, kept stable by something like an autonomic nervous system.[^ch06-beer-body] The bodies of the people who staff it are below the line, and the picture does not reach them.

An organization, however, is enacted by bodies. It is carried by people whose knowledge is uneven and partly tacit, whose health varies, and whose time is finite and contested; people who fear specific colleagues and specific consequences, who have an appetite for some kinds of work and not others, and who have will. Everything the model describes actually runs on that substrate, so these facts are more than decoration on an implementation. Yet nowhere in the diagram does anybody want anything. Nothing corresponds cleanly to the capacity of the people expected to keep the system viable, whether physical, cognitive or emotional. As far as the notation is concerned, channels have capacity and the humans staffing them do not.

The claim needs to be stated precisely. It is a claim about the diagram and the practice by which we teach it, and Beer's books do not support a stronger one. He writes that the users of an operations room are "human beings constrained by their own neurophysiological limitations," whom it would be absurd to flood with data they cannot take in. He notes that divisional managers are "often unwilling" to learn the metasystem's language, and calls participation a matter of "the psychology of the system (and not its logic)." In his late account System Five supplies will and judgment, and the Chilean algedonic meter let a person signal a satisfaction "that does not have to be 'rationalized'."[^ch06-beer-human] Then, in the workbook written for diagnosis, he sets all of this aside on purpose. *Diagnosing the System* leaves out psychology, social anthropology and industrial relations, although "a lot is implicit in this book that derives from these fields."[^ch06-beer-dts] Human capacity and will appear in the books in passing, and nowhere in the five-system notation or in the diagnostic method. That is the gap we inherited, and we taught it as if it were the model.

## Meaning named, not modeled

The second form of the charge needs the most care, because Beer's texts say more about meaning than our teaching does.

Variety is a count of distinguishable states. That is what the concept is and what it is good for, and Chapter 3 sets out both the strength and the cost of taking it literally. Organizations, though, do not run on counts. They run on interpretation: on what a signal is taken to mean, by whom, against what history, and with what consequence for the person doing the taking. A figure that arrives in a monthly report is a state of the world already read by somebody with an interest in how it reads.

Three bodies of work locate the missing layer with some care. Bateson called it the difference that makes a difference, which is precisely the distinction between a difference that is merely available and one that is registered as mattering. Vickers located it in appreciative judgment, and argued against the whole regulatory framing that our tradition adopted. On his account, what an institution mostly does is maintain and revise the standards by which it reads its own situation, rather than correct deviations from a set point. Luhmann made meaning constitutive of social systems rather than incidental to them.[^ch06-meaning] Whatever we make of any one of these, together they mark out territory that Beer names and does not model.

He names it more often than we remember. "A system is not something given in nature, but something defined by intelligence." Whether a change counts as a state of the system depends on whether it "serves, or has no bearing on, the purposes of the system," and "everything depends on how the system is defined, and this in turn depends on who is the observer." Information "is what changes us"; data are its raw material, and "Data of themselves have no value." At each boundary "the 'language' of each subsystem is unique to it," so messages have to be translated. In *Beyond Dispute* (1994), "it is the interpretation of data that procures purpose."[^ch06-beer-meaning] Beer's sentence on information does the work of Bateson's difference that makes a difference.

What Beer states in principle, the notation does not carry. The five systems have channels and capacities and no place for interpretation, which is needed constantly, at every point where something is declared relevant, out of range or worth escalating. The diagnostic method starts from an act of interpretation, "Choose and clearly define the System-in-focus,"[^ch06-beer-focus] and gives no account of how that choice, or any later reading, is made. This is the partition problem Chapter 3 raises, in a different guise. Distinctions do not come pre-labeled. Someone decides what counts as a state, what counts as a disturbance, and what a given reading means, and Beer says so. The diagram drawn from his model begins after that decision has been made and presents the result as structure. In our teaching, the point at which the observer entered drops out of view.

Two failures the book treats separately connect here. Chapter 2 argues that System Four is not a pair of eyes, partly because scanning presupposes a prior settlement about what is worth looking at. Chapter 25 asks what a transducer actually does when a signal crosses a boundary; on our reading, most of the work it is credited with is interpretive work we have not modeled.

## Identity at the top compounds it

Placing identity at the apex of the diagram makes the meaning problem worse. It offers the analyst a resting place. Coordinated action can be explained by installing a smaller interpreter inside the organization, something that already knows what the whole is for and what its situation means, and whose knowing is then transmitted downward. This is the homunculus move. It explains nothing, because the interpreter's own interpreting is exactly what needed accounting for.[^ch06-kenny]

Beer's texts resist this reading more than our diagrams do. System Five "consists of many people," an "elaborately interactive assemblage of elements." Its main problem, he writes, "is its belief that it is *not* the puppet of some other system … But it always is," because its input is mediated by Systems Three and Four, and "*the senior controllers inevitably work within the framework set by their juniors*." It sets the rules "not so much by stating them firmly, as by creating a corporate ethos," and Beer disowns the "connotations of seniority and autocracy" that drawing the metasystem on top carries.[^ch06-beer-s5] Against this stands the language of *Brain of the Firm*, where System Five is "the thinking part … the cortex itself" and a "thinking chamber" originates intentions.[^ch06-beer-cortex] On balance the objection lands on our practice and on that cortex language, and not on the System Five of *The Heart of Enterprise* and *Diagnosing the System*. What remains open is the specialist literature: we have not identified a statement of the homunculus objection against the VSM in the form we would need.

## What would settle it

As it stands, the charge is a reading, and readings are cheap. Chapter 9 explains why our tradition accumulated so many of them: nothing we asserted was ever priced, so nothing was ever tested. This one can be priced.

**Receipt — one comparison paper.** The test is a systematic mapping of each of the five systems against the commitments of 4E cognition[^ch06-4e] — embodied, embedded, enacted, extended — stating for each component what changes under those commitments, what survives them intact, and what cannot be translated at all. Price: one focused review paper. The negative result must be allowed to count. If a faithful mapping leaves every component intact, the negative result defeats the disembodiment charge and the charge should be withdrawn.

We would regard that outcome as a good one. It would mean the model is more robust than we suspect, and that the resources for handling embodiment and meaning were present all along in a form we had failed to teach. Either result advances the position; only the absence of the test leaves us where we are. Chapter 28 sets out the general form of this commitment. The specific form here is simple: a criticism we are unwilling to lose is not a criticism. It is an allegiance.

Two limits remain. First, the mapping we are asking for is comparative, not confirmatory. A paper that sets out to show the VSM was 4E all along will not do the job, because the translation failures are the informative part. Second, this chapter has more unverified ground than most in Part One, and a reader is entitled to hold that against us. Chapter 7 argues that the model never became the machine it was described as being. The argument here is narrower and, for now, weaker: Beer wrote about bodies and readers, the notation he left us acquired neither, and we know exactly what it would cost to find out whether it can.

[^ch06-meaning]: Gregory Bateson, *Steps to an Ecology of Mind* (Chandler Publishing, 1972), for the difference that makes a difference; Geoffrey Vickers, *The Art of Judgment: A Study of Policy Making* (Chapman & Hall, 1965), for appreciative judgment and the standards by which an institution reads its own situation; Niklas Luhmann, *Social Systems*, trans. John Bednarz Jr. with Dirk Baecker (Stanford University Press, 1995), chap. 2, "Meaning," and *Organization and Decision*, ed. Dirk Baecker, trans. Rhodes Barrett (Cambridge University Press, 2018), for meaning as constitutive of social systems.

[^ch06-vat]: Hilary Putnam, *Reason, Truth and History* (Cambridge University Press, 1981), chap. 1, "Brains in a Vat."

[^ch06-4e]: Albert Newen, Leon De Bruin, and Shaun Gallagher, eds., *The Oxford Handbook of 4E Cognition* (Oxford University Press, 2018), https://doi.org/10.1093/oxfordhb/9780198735410.001.0001.

[^ch06-teaching]: Our description of our own teaching (Krishan's and Meg's course material); compare the diagrams in Stafford Beer, *Diagnosing the System for Organizations* (John Wiley & Sons, 1985), and in Jon Walker, *The Viable Systems Model Guide*, 3rd ed., version 3.2 (2020), https://vsmg.lrc.org.uk/, cited in Chapter 1.

[^ch06-kenny]: For the general form of the objection, Anthony Kenny, "The Homunculus Fallacy," in Marjorie Grene, ed., *Interpretations of Life and Mind: Essays around the Problem of Reduction* (Routledge & Kegan Paul, 1971). We have not yet read it; pages will follow when we have.

[^ch06-beer-body]: Stafford Beer, *Brain of the Firm*, 2nd ed. (John Wiley & Sons, 1981), 75–76, on the heart and liver as the example of necessary autonomy, and 103 on the autonomic system that keeps "a stable internal environment."

[^ch06-beer-human]: Stafford Beer, *The Heart of Enterprise* (John Wiley & Sons, 1979), 243; Beer, *Brain of the Firm*, 159 and 283; Stafford Beer, *Think Before You Think*, ed. David Whittaker (Wavestone Press, 2009), 150, for System Five as will and judgment; and *Heart of Enterprise*, 91, where "human beings become tired or distracted. At those moments, requisite variety is lost." The manager's "psychological space" and perceptual filter are in *Heart of Enterprise*, 381–93.

[^ch06-beer-dts]: Stafford Beer, *Diagnosing the System for Organizations* (John Wiley & Sons, 1985), ii.

[^ch06-beer-meaning]: Stafford Beer, *Decision and Control* (John Wiley & Sons, 1966), 242 and 274; Beer, *Diagnosing the System*, 98; Stafford Beer, *Platform for Change* (John Wiley & Sons, 1975), 223; Beer, *Heart of Enterprise*, 101; Stafford Beer, *Beyond Dispute: The Invention of Team Syntegrity* (John Wiley & Sons, 1994), 10.

[^ch06-beer-focus]: Beer, *Diagnosing the System*, 16.

[^ch06-beer-s5]: Beer, *Heart of Enterprise*, 263–64 and 200; Beer, *Brain of the Firm*, 205 and 238; Beer, *Diagnosing the System*, 124.

[^ch06-beer-cortex]: Beer, *Brain of the Firm*, 201 and 135.